Clock recovery units are normally connected to a data source and recover a clock rate characteristic of the data source. Normally such clock recovery units are operated using phase locked loops (PLL).
In doing so, the clock rate via which the data streams in question are sent is recovered. Duplicate networks are used in safety-critical applications. For safety reasons, different clock recovery units are used for such duplicate networks. Data is then transmitted using two separate systems which exchange their data.
Data processing systems which process a plurality of data streams may be driven by the same clock signal. Since a common clock signal then processes the at least two data streams and since this clock signal is recovered from the data streams using a clock recovery unit having a phase locked loop, it is necessary for the clock signal to be available or recoverable if one of the at least two data streams is lost due to an error in the data system.
In order to solve this problem, conventional data processing systems use switching devices which ensure that a clock recovery is shifted to the at least one second data stream by the clock recovery unit if one data stream is lost. In a disadvantageous manner, however, a phase shift generally exists between the first data stream and the at least one data stream in such a way that a settling time is necessary if the clock recovery unit recovers and synchronizes a clock signal from another data stream. Such settling times may be in the order of magnitude of 10 to 20 milliseconds (ms); such a settling time is not tolerable for safety-critical applications.
For safety reasons, clock recovery units of this type are not usable for x-by-wire systems in particular.